Anda di halaman 1dari 8

Protein Engineering vol.14 no.6 pp.

447454, 2001

Secreted production of a custom-designed, highly hydrophilic


gelatin in Pichia pastoris

Marc W.T.Werten1, Wouter H.Wisselink, Tanja J.Jansenvan den Bosch2, Eric C.de Bruin and Frits A.de Wolf
Agrotechnological Research Institute (ATO BV), Bornsesteeg 59, 6708 PD
Wageningen, The Netherlands
2Present

address: NIZO Food Research, Kernhemseweg 2, 6718 ZB Ede,


The Netherlands
1To

whom correspondence should be addressed.


E-mail: m.w.t.werten@ato.wag-ur.nl

A custom-designed, highly hydrophilic gelatin was produced in Pichia pastoris. Secreted production levels in
single-copy transformants were in the range 36 g/l of
clarified broth and purification to near homogeneity could
be accomplished by differential ammonium sulfate precipitation. Despite the fact that gelatins are highly susceptible
to proteolysis because of their unfolded structure, the
recombinant protein was shown to be fully intact by
SDSPAGE, N-terminal sequencing, gel filtration chromatography and mass spectrometry. Owing to its highly
hydrophilic nature, the migration of the synthetic gelatin
in SDSPAGE was severely delayed. Esterification of the
carboxylic amino acid side chains resulted in normal
migration. The high polarity of the synthetic gelatin also
accounts for its negligible surface activity in water at
concentrations up to 5% (w/v), as determined by tensiometry. Circular dichroism spectrometry showed that the
non-hydroxylated gelatin did not form triple helices at 4C.
The spectrum was even more representative of the random
coil conformation than the spectrum of natural nonhydroxylated gelatins.
Keywords: collagen-like protein/hydrophilic gelatin/Pichia
pastoris/proteolytic stability/synthetic gene

Introduction
Gelatin, in essence denatured and partially degraded collagen,
is traditionally prepared by hot acid or alkaline extraction of
animal bones and hides. Apart from its main use as a gelling
agent in food (Asghar and Henrickson, 1982), gelatin is also
used in medical and industrial applications such as intravenous
infusions (Saddler and Horsey, 1987), matrix implants (Pollack,
1990), injectable drug delivery microspheres (Rao, 1995) and
photographic film (Courts, 1980). Despite the diversity of
current uses of natural gelatin, recombinant gelatins may
provide benefits for specific applications, in that the chemical
composition and molecular weight can be precisely controlled
and reproduced (van Heerde et al., 1999; de Wolf et al., 2000).
Furthermore, recombinant gelatins do not bear the risk of
associated infectious diseases such as bovine spongiform
encephalopathy (BSE).
Several reports have described the production of recombinant
gelatin-like proteins in Escherichia coli. Analogously to the
natural amino acid sequence of the collagen triple-helix forming
domain, synthetic genes are constructed from repeating (Gly
Oxford University Press

XaaYaa)n-encoding oligonucleotides, where Xaa and Yaa are


often proline (Goldberg et al., 1989; Obrecht et al., 1991;
Gardner et al., 1993; Cappello and Ferrari, 1994). Gene
instability problems are commonly observed with such highly
repetitive genes (Capello and Ferrari, 1994). Also, expression
levels usually obtained in E.coli are rather low and purification
of the intracellularly produced protein can be difficult. Recently,
Kajino et al. reported the use of Bacillus brevis for the
expression of gelatin-like proteins (Kajino et al., 2000). They
used sequence stretches selected from natural collagen genes
and polymerized them to form semi-synthetic gelatin. Prior to
their report, we reported the use of the methylotrophic yeast
Pichia pastoris as a superior host for the secretion of recombinant gelatins having natural amino acid sequences, at up to
14.8 g/l of clarified broth (Werten et al., 1999).
Having established the suitability of P.pastoris for the
expression of natural recombinant gelatins, we set out to
investigate the possibilities of producing entirely customdesigned gelatins having novel physico-chemical properties.
A monomeric gene encoding a highly hydrophilic 9 kDa gelatin
was designed such as to allow convenient polymerization into
larger multimers. The monomeric gene is much longer than
the single oligonucleotide monomers used in the expression
of synthetic gelatins in E.coli, mentioned above. This offers
more flexibility in the design of the amino acid sequence and
a concomitant decrease in the overall repetitiveness of the
gene. Here, we describe the high-level secretion of a fully
synthetic, highly hydrophilic and non-degraded 36.8 kDa
gelatin by P.pastoris and its characterization.
Materials and methods
Vector construction
The monomeric gelatin gene (referred to hereafter as P for
polar) was constructed by overlap extension polymerase
chain reaction (PCR) (Ho et al., 1989) of long oligonucleotides
(underlined in Figure 1A). PCR was performed with a PerkinElmer GeneAmp 9700, using the proofreading enzyme Pwo
DNA polymerase (Eurogentec). The 5 half of the gene was
constructed by overlap extension of the first and second
oligonucleotides and co-amplified by outer primers directed
against nucleotides 126 (sense) and 197174 (antisense).
Likewise, the 3 half of the gene was constructed by overlap
extension of the third and fourth oligonucleotides and coamplified by primers directed against nucleotides 174197
(sense) and 363337 (antisense). The resulting PCR products
were isolated from an agarose gel and were combined by
another overlap extension PCR and co-amplified with the
primers directed against nucleotides 126 (sense) and 363
337 (antisense). The resulting 0.3 kb PCR fragment was
digested with XhoI/EcoRI and cloned in vector pMTL23
(Chambers et al., 1988) to form vector pMTL23P. The sequence
of the gene was verified by automated DNA sequencing of
both strands.
The monomeric gene was released by digesting pMTL23P
447

M.W.T.Werten et al.

Fig. 1. Construction of the synthetic gelatin gene. (A) Overlapping oligonucleotides and encoded P monomer. Letters in bold type indicate the non-ambiguous
nucleotides recognized by DraIII and Van91I. (B) Multimerization of P monomer to P4. Note that Van91I/DraIII hybrid sites are not recleavable.

with DraIII/Van91I (Figure 1B). In a separate reaction the


vector was linearized with Van91I and dephosphorylated. The
DraIII/Van91I fragment was then inserted into this linearized
vector to yield vector pMTL23P2. This process of insertional
448

doubling can in principle be repeated to form multimers of


any desired length, but was repeated only once here to
form the vector pMTL23P4. The tetrameric gene (referred to
hereafter as P4) was then cloned into the XhoI/EcoRI sites

Secreted production of hydrophilic gelatin in P.pastoris

of vector pPIC9 (Invitrogen) that contains a HIS4 selectable


marker, an alcohol oxidase 1 (AOX1) promoter/terminator
cassette and a Saccharomyces cerevisiae -factor prepro secretory signal (Clare et al., 1991a). The Kex2 and dipeptidylaminopeptidase (DPAPase) cleavage sites of the -factor prepro
sequence are lost from pPIC9 when using the XhoI site, but
are restored upon ligation of the gelatin gene by the sequence
between XhoI and DraIII (Figure 1A).
Transformation of P.pastoris
Plasmid pPIC9P4 was linearized with SalI in order to obtain
preferentially Mut transformants [i.e. by integration at the
his4 locus rather than the AOX1 locus and thus allowing normal
growth on methanol (Clare et al., 1991b)]. Transformation
of P.pastoris strain GS115 [his4 (Cregg et al., 1985)] by
electroporation and selection of Mut transformants was as
described previously (Werten et al., 1999).
Fermentative production of synthetic gelatin in P.pastoris
Fermentations were performed in 1140 l fermenters
(Applikon) in minimal basal salt medium (Invitrogen)
supplemented with 0.2% (v/v) PTM1 trace salts (Invitrogen).
Methanol fed-batch fermentations were performed as described
previously (Werten et al., 1999), with the exception that no
protease-inhibiting supplements such as casamino acids were
added and that the pH during methanol fed-batch was maintained at 3.0 for all fermentations.
Small-scale purification of synthetic gelatin by differential
acetone precipitation
Differential acetone precipitation was as described previously
(Werten et al., 1999). Chilled acetone was added to fermentation supernatant at 40% (v/v), after which endogenous proteins
were pelleted by centrifugation. The acetone concentration in
the supernatant was then increased to 80% (v/v) and the pellet
obtained after centrifugation was washed with 80% acetone
and air-dried.
Preparative purification of synthetic gelatin by differential
ammonium sulfate precipitation
Preparative purification of synthetic gelatin from fermentation
supernatant consisted of twice-repeated ammonium sulfate
precipitation at 40% saturation (4C) and subsequent washing
of the precipitate with 60% saturated ammonium sulfate.
Depending on the scale of the purification, separation of the
precipitate from the liquid was either by centrifugation or by
depth filtration using AKS-4 sheets (USF Seitz-Schenk). The
protein was subsequently desalted by diafiltration and lyophilized.
Bicinchoninic acid protein assay
A commercially available bicinchoninic acid (BCA) protein
assay was used according to the manufacturers recommendations (Pierce). The reaction was performed at 60C for 30 min.
The calibration curve was prepared gravimetrically from lyophilized, desalted P4 gelatin, purified by differential ammonium
sulfate precipitation (purity at least 98%).
SDSPAGE
SDSPAGE (Laemmli, 1970) was performed in a MiniPROTEAN II system (Bio-Rad) under reducing denaturing
conditions. Gels consisted of a 5% stacking and a 12.5%
separating zone (2.7% cross-linking). Gels were stained using
Coomassie PhastGel Blue R-350 (Amersham-Pharmacia
Biotech) and were destained by heating in water using a
microwave oven, similarly to Faguy et al. (Faguy et al., 1996).

Gel filtration chromatography


Protein in 0.1 M sodium chloride was loaded on a 10250 mm
column packed with Superose 12 (Amersham-Pharmacia
Biotech). Elution was carried out with 0.1 M sodium chloride
at a flow-rate of 0.2 ml/min, collecting 2 ml fractions and
monitoring the absorbance at 214 nm.
Mass spectrometry
Matrix-assisted laser desorption/ionization time-of-flight mass
spectrometry (MALDI-TOF MS) was performed at the
Department of Biochemistry, Wageningen University, The
Netherlands, using a Voyager DE-RP delayed extraction mass
spectrometer (PerSeptive Biosystems). Samples were prepared
by the dried droplet method, using sinapinic acid dissolved in
30% (v/v) acetonitrile, 4% (v/v) trifluoroacetic acid as matrix.
Measurements were made in the positive, linear mode and the
accelerating voltage was 25 000 V. Cytochrome c and bovine
serum albumin were used as external calibrants.
Chemical modification of gelatins
Esterification of carboxylic amino acid side chains was adapted
from Wilcox (Wilcox, 1967). A 100 g amount of protein was
incubated in 500 l of methanol, 0.1 M hydrochloric acid at
4C for 72 h. The methanol was exchanged for 1 mM
hydrochloric acid by diafiltration in a 3 kDa Microcon (Millipore). Removal of ester groups was performed by incubating
the esterified protein in 100 mM TrisHCl, pH 8.8 at 20C
for 72 h.
Hydrazination of carboxylic amino acid side chains was
performed according to Matagne et al. (Matagne et al., 1991).
A 50 g amount of protein was dissolved in 20 l of 50 mM
sodium phosphate buffer, pH 7. After addition of 170 l of 8
M urea, 1 M hydrazine and 0.1 M 1-(3-dimethylaminopropyl)3-ethylcarbodiimide, pH 4.5, the mixture was incubated at
room temperature for 2 h.
Surface tension measurements
Surface tension at the liquidair interface was measured
according to the du Nou y ring method (Lecomte du Nou y,
1919) using a Kru ss K6 tensiometer. The temperature of the
sample vessel was maintained at 20C. The actual measurements were performed 5 min after lowering of the ring on to
the liquid surface, as suggested by Clarkson et al. (Clarkson
et al., 1999). Bovine serum albumin was used as a reference
protein and raw data were corrected for the hydrostatic
volume effect according to Harkins and Jordan (Harkins and
Jordan, 1930).
Circular dichroism spectrometry
Proteins were dissolved in Milli-Q water at 0.1 mg/ml. Measurements were performed at the Department of Biochemistry,
Wageningen University, The Netherlands using a Jasco J-715
spectropolarimeter. The pathlength was 0.1 cm and spectra
were recorded from 190 to 260 nm at 4C, using a scanning
speed of 20 nm/min at a resolution of 0.1 nm.
Results
Design and construction of the synthetic P4 gelatin gene
The basic structure of natural gelatins consists of repeating
GlyXaaYaa triplets, where Xaa and Yaa are often proline
and hydroxyproline, respectively (the latter being posttranslationally modified proline). This structure maintains the
open, unfolded conformation characteristic of gelatin. Owing
to this unfolded conformation, gelatin is fairly hydrophilic
449

M.W.T.Werten et al.

Table I. Composition and basic physico-chemical parameters of synthetic


and natural gelatinsa
Amino acid

P4

Ala
Arg
Asn
Asp
Asx
Cys
Gln
Glu
Glx
Gly
His
Ile
Leu
Lys
Met
Phe
Pro
Ser
Thr
Trp
Tyr
Val

1.0
0.0
12.0
0.0

0.0
16.0
4.0

33.7
0.0
0.0
0.0
3.0
0.0
0.0
22.4
8.0
0.0
0.0
0.0
0.0

GRAVY
Isoelectric point
Molecular weight (kDa)

1.77
4.9
36.8

Col3a1b
7.9
4.8
3.5
3.1

Cattle bone gelatinc

0.0
3.5
3.5

33.8
1.8
1.3
1.8
3.5
0.4
0.4
21.1
6.1
1.8
0.0
0.4
1.3

12.0
5.0

4.3d
0.0

6.7d
33.4
0.5
1.2
2.4
2.6e
0.7
1.3
22.2e
3.4
1.6
0.0
0.2
2.4

1.08
9.7
20.6

0.75 to 1.09e
4.7 to 5.4f
65 to 300f

aTheoretical

amino acid composition (mol%), GRAVY values, isoelectric


point and molecular weight were calculated using the ProtParam tool
available at the Expasy WWW server (Appel et al., 1994). Experimental
values were obtained as indicated below.
bNatural non-hydroxylated recombinant gelatin produced in P.pastoris
(Werten et al., 1999).
cThe amino acid composition (mol%) of limed cattle bone gelatin is the
mean of four measurements.
dAsn/Asp and Gln/Glu cannot be distinguished by amino acid analysis and
are given as Asx and Glx, respectively. This does not affect the GRAVY
calculation.
eValues for Lys and Pro include hydroxylysine and hydroxyproline,
respectively. The hydropathy indices of these modified amino acids were
assumed to lie between those of the respective unmodified amino acids and
the lowest value of the Kyte and Doolittle hydropathy scale. The range of
GRAVY values indicated for cattle bone gelatin represents both extremes.
fNatural cattle bone gelatin is a heterogeneous mixture of molecules of
different molecular weights and isoelectric points within the indicated
ranges (Alleavitch et al., 1988).

because its hydrogen bonds are highly exposed. Furthermore,


only a small fraction of the protein is occupied by hydrophobic
amino acids such as Trp, Tyr, Phe, Leu, Ile, Val and Met.
Our synthetic P4 gelatin design also provides the (Gly
XaaYaa)n structure. To increase its hydrophilicity relative to
that of natural gelatins, we designed a gelatin without any
hydrophobic amino acids other than proline and with a high
content of the hydrophilic amino acids asparagine and glutamine (Table I). To illustrate the high hydrophilicity of this
synthetic gelatin compared with natural gelatins, the GRAVY
values [grand average of hydropathy (Kyte and Doolittle,
1982)] of P4, natural recombinant Col3a1 gelatin (Werten
et al., 1999) and cattle bone gelatin are indicated in Table I.
The content of acidic and basic residues was modulated to
give an isoelectric point similar to that of common limed bone
gelatin (Table I). Only Lys was used as a basic residue, because
we anticipated proteolysis of mono-arginylic sites based on
previous work (Werten et al., 1999).
The P4 gene was constructed from four identical P monomers
450

Fig. 2. Purification of synthetic gelatin. Lane 1, fermentation supernatant;


lane 2, precipitate obtained at 40% (v/v) acetone; lane 3, precipitate
obtained by increasing the acetone concentration of the 40% (v/v)
supernatant to 80% (v/v); lane 4, precipitate obtained at 40% of ammonium
sulfate saturation; lane 5, supernatant obtained at 40% of ammonium sulfate
saturation; lane M, molecular weight marker.

that were designed to have the codon usage of P.pastoris


highly expressed genes (Sreekrishna and Kropp, 1996). The
monomeric gene contains restriction sites for DraIII and
Van91I. These enzymes allow the design of mutually complementary, non-palindromic overhangs that enable convenient
elongation of the gene by insertional doubling in a fixed
orientation (Figure 1). The process can be repeated until the
desired polymer length has been achieved. This modular design
offers flexibility in the construction of future gelatins by
allowing the combination of different types and lengths of
polymerized gelatins via the DraIII/Van91I sites. The XhoI
and EcoRI sites provided by the sequence allow the direct
insertion of the final synthetic gene into P.pastoris expression
vector pPIC9, resulting in a fusion to the alpha-mating factor
prepro secretion signal. Thus, the combined four P modules
were cloned into pPIC9 to yield vector pPIC9P4.
Production of synthetic P4 gelatin
Plasmid pPIC9P4 was used to transform P.pastoris GS115.
Randomly chosen transformants were fermented and culture
supernatants harvested throughout the fermentation were subjected to SDSPAGE. The theoretical molecular weight of P4
is 36.8 kDa. Because collagenous proteins migrate in SDS
PAGE at an apparent molecular weight ~40% higher than the
true molecular weight (Butkowski et al., 1982; Werten et al.,
1999), one would expect a band of ~52 kDa. Instead, however,
the Coomassie Blue-stained SDSPAGE gel showed a faint
blurry band at the top of the separating gel that had a tendency
to diffuse from the gel during methanolacetic acid destaining.
Migration of the gelatin into the gel was improved by running
the gel at 4C at twice the voltage recommended by the
manufacturer of the electrophoresis system (i.e. 400 instead
of 200 V). Diffusion of the protein from the gel during
destaining was reduced by destaining the gel in water heated
in a microwave oven [similarly to Faguy et al. (Faguy et al.,
1996)], rather than performing the common lengthy incubations
in methanolacetic acid. Figure 2, lane 1 shows fermentation
supernatant analyzed in this manner. N-Terminal protein
sequencing of this band (Sequencing Centre Utrecht, The
Netherlands) revealed the expected amino acid sequence
(GPPGEPGNPG). There was no indication of incomplete
processing of the -factor derived GluAla repeats by dipeptidylaminopeptidase, such as is occasionally observed when using
this prepro sequence for secretion of heterologous proteins
(Vedvick et al., 1991; Briand et al., 1999; Werten et al., 1999;
Goda et al., 2000).

Secreted production of hydrophilic gelatin in P.pastoris

Recombinant gelatins with natural amino acid sequences


can be purified from the fermentation broth by using differential
acetone precipitation (Werten et al., 1999). Endogenous extracellular proteins are precipitated at 40% (v/v) acetone, after
which the gelatin is precipitated from the supernatant by
addition of acetone to 80% (v/v). Figure 2, lanes 2 and 3
show that acetone precipitation was equally effective for the
purification of synthetic gelatins. Amino acid analysis was
performed (in triplicate) to estimate the purity of the precipitated gelatin. By linear least-squares fitting of the observed
data in terms of the P4 amino acid composition, the average
relative contributions (vector moduli) of the P4 component
and the residuals component were found to be 96.2 and 3.8%,
respectively (1.4% SD). The purity of the preparation at the
protein level is thus estimated to be 96%, as stochastic
fluctuations of measured values contribute to the residuals
component. BCA protein analysis of acetone precipitates of
different fermentations showed gelatin yields in the range
36 g/l of clarified broth.
We previously found that recombinant gelatins with natural
amino acid sequences can also be purified from P.pastoris
fermentations using differential ammonium sulfate precipitation (unpublished data). Gelatinous proteins precipitate at
40% saturation, whereas endogenous extracellular P.pastoris
proteins surprisingly do not precipitate at up to 80% saturation.
In agreement with this, readily precipitable proteins such as
-lactoglobulin are rendered virtually unprecipitable upon
mixing them with fermentation supernatant. We investigated
whether it was possible to purify P4 gelatin by differential
ammonium sulfate precipitation. Indeed, Figure 2, lanes 4 and
5 show that synthetic gelatin is quantitatively precipitated at
40% ammonium sulfate saturation, while no endogenous
proteins are visible. Based on amino acid analysis (in triplicate)
and subsequent linear least-squares fitting of the observed data,
the purity at the protein level was estimated to be 98.1%
(0.7% SD). Two-dimensional electrophoresis followed by
silver staining showed virtually no contaminants (not shown).
Gelatin yields determined by BCA analysis of ammonium
sulfate precipitates of several fermentations [possible only
after thorough desalting because ammonium sulfate reduces
BCA reactivity (Smith et al., 1985)] were within 0.3 g/l of
the values determined by acetone precipitation. This small
difference (10%) is largely due to interference of the BCA
assay by the small amount of reducing exopolysaccharides
that co-purifies in the acetone precipitation procedure, while
exopolysaccharides are virtually eliminated when using ammonium sulfate precipitation (data not shown). Compared with
differential acetone precipitation, the overall higher purity
obtained and the higher amenability to scale-up render differential ammonium sulfate precipitation the method of choice for
preparative purification of P4 gelatin.
Establishing the molecular weight of synthetic P4 gelatin
A possible explanation for the aberrant molecular weight
observed in SDSPAGE could be that P4 is glycosylated.
N-Linked glycosylation can be ruled out because no susceptible
sites are present in the amino acid sequence. However,
P.pastoris is also able to perform O-glycosylation and the
structural determinants for such an event are unclear (Duman
et al., 1998). To rule out this possibility, periodic acidSchiff
staining (Zacharius et al., 1969) and Alcian Blue staining
(Wardi and Michos, 1972) were performed on ammonium
sulfate-purified P4. No glycosylation was observed (not
shown).

Fig. 3. MALDI-TOF mass spectrum of synthetic gelatin purified by


differential ammonium sulfate precipitation. The [M H] and
[M 2H]2 peaks correspond to singly and doubly charged molecular ions,
respectively.

To determine whether the synthetic gelatins have in fact the


correct molecular weight, but merely exhibit aberrant behavior
in SDSPAGE, analytical gel filtration chromatography was
performed. The Superose 12 column was calibrated with a
mixture of natural recombinant gelatin fragments (Werten
et al., 1999), giving a series of molecular weights of 53, 42,
28, 16, 12 and 8 kDa. Ammonium sulfate-purified P4 was
subjected to gel filtration chromatography. Only one significant
peak was observed. N-Terminal protein sequencing of this
fraction in solution (Sequencing Centre Utrecht) showed the
correct N-terminus for P4. The molecular weight of P4 deduced
from the chromatogram was 47 kDa. This is clearly much
closer to the theoretical value of 36 kDa than the molecular
weight apparent from SDSPAGE, although the deviation is
still significant.
Mass spectrometry was used to determine ultimately the
molecular weight of P4. Materials purified by both ammonium
sulfate precipitation and gel filtration chromatography were
analyzed and the results were in good mutual agreement.
Figure 3 shows the MALDI-TOF mass spectrum of P4 purified
by ammonium sulfate precipitation. The observed molecular
weight of 36 835 Da corresponds well with the theoretical
value of 36 818 Da. This result shows that the apparent high
molecular weight observed in SDSPAGE is indeed the result
of aberrant migration behavior. Furthermore, the SDSPAGE
and gel filtration chromatography results are confirmed, in that
there is no presence of proteolytically degraded fragments.
We speculated that the aberrant migration rate of synthetic
gelatin compared with normal gelatin was due to low binding
of SDS in view of its high hydrophilicity, as the interaction
of SDS with proteins is mainly of a hydrophobic nature
(Reynolds and Tanford, 1970). Esterification of the carboxylic
amino acid side chains with methanolhydrochloric acid would
increase the proteins hydrophobicity and thus its SDS binding
capacity and migration rate. While the migration of natural
gelatins treated in this way was only slightly affected (i.e. a
decrease in apparent molecular weight of about 2 kDa), the
esterified synthetic gelatin migrated much faster than the
unmodified protein (Figure 4). The molecular weight observed
for esterified P4 was about 50 kDa. As natural gelatins migrate
about 40% more slowly than common proteins (Butkowski
et al., 1982; Werten et al., 1999), this value is in good
agreement with the value expected for natural gelatins of the
same molecular weight as P4 (i.e. a 36.8 kDa natural gelatin
would migrate at ~52 kDa). Removal of the ester groups at
high pH restored the aberrant migration of P4 (Figure 4).
Hydrazination of the carboxylic amino acid side chains of P4
did not result in an altered migration rate (not shown),
451

M.W.T.Werten et al.

Fig. 4. Esterification of natural recombinant Col3a1 and synthetic P4


gelatin. Lane 1, unmodified Col3a1 natural recombinant gelatin; lane 2,
esterified Col3a1; lane 3, esterified Col3a1 after hydrolysis of the ester
groups; lane 4, unmodified P4; lane 5, esterified P4; lane 6, esterified P4
after hydrolysis of the ester groups; lane M, molecular weight marker.

Fig. 6. Circular dichroism spectra of natural recombinant Col3a1 and


synthetic P4 gelatin at 4C.

helix formation. Figure 6 shows the spectrum of P4 and natural


recombinant Col3a1 gelatin at 4C. Both proteins show a clear
absence of the positive peak at about 221 nm characteristic of
the collagen triple helix (de Wolf and Keller, 1996; Rossi
et al., 1996). The mean residue ellipticity of P4 at that position
of the spectrum is lower than that of Col3a1.

Fig. 5. Surface tension of natural recombinant Col3a1 and synthetic P4


gelatin as a function of concentration. Error bars indicate the standard
deviation of five measurements. The dotted line indicates the surface tension
of water.

indicating that the effect of esterification was not due to the


mere removal of negative charge.
Characterization of synthetic P4 gelatin
In marked contrast with recombinant gelatins having natural
sequences, we noticed that highly concentrated solutions of
P4 showed essentially no foaming. A direct relationship exists
between protein (surface) hydrophobicity, surface tension and
foam stability (Horiuchi et al., 1978). Therefore, the surface
activity of P4 relative to that of Col3a1 natural recombinant
gelatin was determined, using the du Nou y ring method
(Lecomte du Nou y, 1919). Figure 5 shows that P4 does not
show any significant lowering of the surface tension of water
at concentrations up to 5% (w/v), whereas Col3a1 already has
an effect at 0.01%. Within the range up to 10% of protein, it
was not possible to determine the apparent critical micelle
concentration (CMC; i.e. the concentration whereby the surface
tension curve reaches a plateau phase) for either of the gelatin
types. For comparison, bovine serum albumin has an apparent
CMC of about 0.003% (Clarkson et al., 1999).
We previously showed that natural, non-hydroxylated recombinant gelatins produced in P.pastoris, do not form collagen
triple helices even at 4C (Werten et al., 1999). Synthetic P4
gelatin was also non-hydroxylated, as shown by the amino
acid analyses and N-terminal sequencing described above.
Circular dichroism spectrometry was performed to see whether
the greater hydrophilicity of P4 somehow influences triple452

Discussion
This paper describes the extracellular production of a recombinant, custom-designed and highly hydrophilic gelatin in P.pastoris. Yields in single-copy transformants were 36 g/l of
clarified broth, which is comparable to the yields obtained in
single-copy transformants for natural gelatins (Werten et al.,
1999).
Despite the obvious physico-chemical differences, both
synthetic P4 and natural recombinant gelatin could be purified
from the fermentation broth by both differential acetone
precipitation and differential ammonium sulfate precipitation.
The universality of the two purification techniques is probably
due to the hydrophilicity and unfolded structure of gelatins in
general. Especially differential ammonium sulfate precipitation
allowed convenient large-scale purification of P4 gelatin to
near homogeneity.
Secreted synthetic gelatin was fully intact, as evidenced by
SDSPAGE, gel filtration chromatography, N-terminal sequencing and mass spectrometry. This is in contrast with natural
recombinant gelatins produced in P.pastoris, which were partly
degraded (Werten et al., 1999). Apart from the occurrence of
some minor background degradation, collagen type I-derived
natural recombinant gelatins were cleaved into several major
bands by a Kex2-like protease. Cleavage occurred C-terminal
of two occurrences of the mono-arginylic sequence MetGly
ProArg. We speculated that the amino acids occupying the
2 and 4 positions in this motif (relative to the site of cleavage)
were a major factor determining the cleavage efficiency (Werten
et al., 1999), which is in accordance with recent data on the
substrate specificity of Saccharomyces cerevisiae Kex2 (Bevan
et al., 1998; Suzuki et al., 2000). It may well be that the
above-mentioned minor background degradation represented
a limited extent of cleavage at Arg residues having suboptimal
residues at the 2 and 4 positions. Therefore, in the design
of the synthetic gelatin described here, only Lys was used as
a basic residue to control the isoelectric point. The finding
that secreted P4 was completely intact does indeed suggest a
general susceptibility of Arg residues in recombinant gelatins
to proteolysis and may thus have implications for the rational

Secreted production of hydrophilic gelatin in P.pastoris

design of (partially) unfolded proteins to be expressed extracellularly in P.pastoris.


While the molecular weight of P4 as determined by mass
spectrometry was in good agreement with the value deduced
from the amino acid sequence, the molecular weight apparent
from gelatin-calibrated gel filtration chromatography was about
10 kDa higher. Ionic interactions with Superose 12 are negligible in the presence of salt and only small hydrophobic
peptides appear to show significant hydrophobic interactions
with this matrix (Andersson et al., 1985). It is therefore not
very likely that a lower degree of such interactions of P4
relative to natural gelatins causes the seemingly aberrant
molecular weight. Possibly the effect is due to an increased
hydrodynamic size of P4 relative to natural gelatins, as a result
of increased interaction of this highly hydrophilic protein
with water.
Natural gelatins migrate about 40% more slowly in SDS
PAGE than expected. Several possible explanations for the
aberrant migration behavior of gelatins have been suggested
(Furthmayr and Timpl, 1971; Freytag et al., 1979; Hayashi
and Nagai, 1980; Noelken et al., 1981; Butkowski et al.,
1982). It is most likely not the result of anomalously low SDS
binding, but is at least in part due to the low average residue
molecular weight of gelatin, resulting in a relatively high
number of residues (i.e. molecular length) per unit of molecular
weight (Freytag et al., 1979; Noelken et al., 1981; Butkowski
et al., 1982). Most other reports on aberrant protein migration
rates in SDSPAGE involve highly acidic proteins that show
reduced binding of SDS due to electrostatic repulsion by the
proteins high negative net charge (Ohara and Teraoka, 1987;
Matagne et al., 1991; Casare gola et al., 1992; McGrath et al.,
1992). SDSPAGE showed that P4 migrates at a highly reduced
rate, even much more slowly than natural gelatins. We showed
here that esterification of the carboxylic side chains of P4
restores its migration rate roughly to that expected for normal
gelatins (i.e. about 40% more slowly than common proteins).
In contrast, hydrazination did not affect the migration rate of
P4. Hydrazination eliminates the negative charge of the same
carboxylic residues as does esterification, but reduces the
proteins hydrophobicity whereas esterification increases it. As
the binding of SDS to proteins is primarily hydrophobic in
nature (Reynolds and Tanford, 1970), the extremely slow
migration of the highly polar P4 gelatin in SDSPAGE is
therefore most likely the result of insufficient SDS binding
and a concomitant low negative net charge. The finding that
the resolution of the SDSPAGE was improved by increasing
the field strength to twice that recommended by the manufacturer of the electrophoresis system indicates that the higher field
strength aids protein migration in overcoming diffusive forces.
Surface activity is a major determinant in a proteins function
as a protective colloid (e.g. in photographic emulsions). Solutions of P4 showed essentially no foaming and tensiometric
analysis of solutions of P4 in water showed only negligible
surface activity. P4 thus represents a novel hydrocolloid
combining some of the characteristics unique to gelatins and
a low surface activity commonly expected only for the most
hydrophilic polysaccharide hydrocolloids.
We previously showed that non-hydroxylated, natural recombinant gelatins do not show triple-helical structure in circular
dichroism spectrometry (Werten et al., 1999). It is a wellestablished fact that hydroxyproline residues play a crucial
role in the stabilization of the collagen triple helix. This
role is easily recognized when examining the amino acid

compositions and thermal stabilities of natural collagens from


different species (Privalov, 1982). X-ray crystallography
showed that the triple helix is surrounded by a cylinder of
hydration (Bella et al., 1994). Although recently questioned
(Holmgren et al., 1999; Nagarajan et al., 1999), the role of
hydroxyproline in the stabilization of the triple helix is generally attributed to its hydrogen bonding with this water network
(Brodsky and Shah, 1995). In view of the high polarity of P4,
we considered it prudent to investigate its conformation using
circular dichroism spectrometry. No triple helical structure was
observed at 4C and P4 gelatin is thus an essentially nongelling gelatin. Non-gelling gelatins permit novel applications
such as low-temperature silver halide crystallization in the
preparation of photographic emulsions (de Wolf et al., 2000).
Comparison of the circular dichroism spectrum of P4 with
that of Col3a1 natural recombinant gelatin showed that the
mean residue ellipticity of P4 at the discriminating wavelength
of 221 nm was actually lower than that of Col3a1. Although
the latter is essentially in a random coil conformation, the higher
hydrophilicity of P4 probably reduces minor intramolecular and
intermolecular interactions, thereby resulting in a slightly lower
ellipticity.
Current research is directed towards the production of other
synthetic gelatins with distinct functionalities and combining
them to form chimeric tailor-made biopolymers.
Acknowledgements
We thank Ir Friso B.J.van Assema, Ir Harald G.de Jonge, Ing Pieter H.Lodder,
Ir Marcia P.M.Dielissen and Ing Marjo Koets for assistance, Ir Harrold
A.van den Burg (Department of Biochemistry, Wageningen University, The
Netherlands) for mass spectrometry, Dr Ir Carlo P.M.van Mierlo (Department
of Biochemistry, Wageningen University, The Netherlands) for help with
circular dichroism measurements and Dr Ir Johan M.H.Stoop for critical
reading of the manuscript.

References
Alleavitch,J., Turner,W.A. and Finch,C.A. (1988) In Ullmanns Encylopedia
of Industrial Chemistry, vol. A12. Wiley-VCH, Weinheim, pp. 307317.
Andersson,T., Carlsson,M., Hagel,L., Pernemalm,P. and Janson,J. (1985)
J. Chromatogr., 326, 3344.
Appel,R.D., Bairoch,A. and Hochstrasser,D.F. (1994) Trends Biochem. Sci.,
19, 258260.
Asghar,A. and Henrickson,R.L. (1982) Adv. Food Res., 28, 231372.
Bella,J., Eaton,M., Brodsky,B. and Berman,H.M. (1994) Science, 266, 7581.
Bevan,A., Brenner,C. and Fuller,R.S. (1998) Proc. Natl Acad. Sci USA, 95,
1038410389.
Briand,L., Perez,V., Huet,J.C., Danty,E., Masson,C. and Pernollet,J.C. (1999)
Protein Express. Purif., 15, 362369.
Brodsky,B. and Shah,N.K. (1995) FASEB J., 9, 15371546.
Butkowski,R.J., Noelken,M.E. and Hudson,B.G. (1982) Methods Enzymol.,
82, 410423.
Capello,J. and Ferrari,F. (1994) In Mobley,D.P. (ed.), Plastics from Microbes.
Carl Hanser, Munich, pp. 3592.
Casare gola,S., Jacq,A., Laoudj,D., McGurk,G., Margarson,S., Tempe te,M.,
Norris,V. and Holland,I.B. (1992) J. Mol. Biol., 228, 3040.
Chambers,S.P., Prior,S.E., Barstow,D.A. and Minton,N.P. (1988) Gene, 68,
139149.
Clare,J.J., Romanos,M.A., Rayment,F.B., Rowedder,J.E., Smith,M.A.,
Payne,M.M., Sreekrishna,K. and Henwood,C.A. (1991a) Gene, 105, 205
212.
Clare,J.J., Rayment,F.B., Ballantine,S.P., Sreekrishna,K. and Romanos,M.A.
(1991b) Biotechnology (NY), 9, 455460.
Clarkson,J.R., Cui,Z.F. and Darton,R.C. (1999) J. Colloid Interface Sci., 215,
333338.
Courts,A. (1980) In Grant B.A. (ed.), Applied Protein Chemistry. Applied
Science, London, pp. 129.
Cregg,J.M., Barringer,K.J., Hessler,A.Y. and Madden,K.R. (1985) Mol. Cell.
Biol., 5, 33763385.

453

M.W.T.Werten et al.
De Wolf,F.A. and Keller,R.C.A. (1996) Prog. Colloid Polym. Sci., 102, 914.
De Wolf,F.A., Werten,M.W.T, Wisselink,H.W., Jansen-van den Bosch,T.J.,
Toda,Y., van Heerde,G.V. and Bouwstra,J.B. (2000) Eur. Pat. Appl. 1014176.
Duman,J.G., Miele,R.G., Liang,H., Grella,D.K., Sim,K.L., Castellino,F.J. and
Bretthauer,R.K. (1998) Biotechnol. Appl. Biochem., 28, 3945.
Faguy,D.M., Bayley,D.P., Kostyukova,A.S., Thomas,N.A. and Jarrell,K.F.
(1996) J. Bacteriol., 178, 902905.
Freytag,J.W., Noelken,M.E. and Hudson,B.G. (1979) Biochemistry, 18,
47614768.
Furthmayr,H. and Timpl,R. (1971) Anal. Biochem., 41, 510516.
Gardner,K., Lock,R.L., OBrien,J.P. and Salemme,F.R. (1993) PCT Pat.
Appl. 931023.
Goda,S., Takano,K., Yamagata,Y., Katakura,Y. and Yutani,K. (2000) Protein
Eng., 13, 299307.
Goldberg,I., Salerno,A.J., Patterson,T. and Williams,J.I. (1989) Gene, 80,
305314.
Harkins,W.D. and Jordan,H.F. (1930) J. Am. Chem. Soc., 52, 17511771.
Hayashi,T. and Nagai,Y. (1980) J. Biochem. (Tokyo), 87, 803808.
Ho,S.N., Hunt,H.D., Horton,R.M., Pullen,J.K. and Pease,L.R. (1989) Gene,
77, 5159.
Holmgren,S.K., Bretscher,L.E., Taylor,K.M. and Raines,R.T. (1999) Chem.
Biol., 6, 6370.
Horiuchi,T., Fukushima,D., Sugimoto,H. and Hattori,T. (1978) Food Chem.,
3, 3542.
Kajino,T., Takahashi,H., Hirai,M. and Yamada,Y. (2000) Appl. Environ.
Microbiol., 66, 304309.
Kyte,J. and Doolittle,R.F. (1982) J. Mol. Biol., 157, 105132.
Laemmli,U.K. (1970) Nature, 227, 680685.
Lecomte du Nou y,P. (1919) J. Gen. Physiol., 1, 521524.
Matagne,A., Joris,B. and Fre`re,J.M. (1991) Biochem. J., 280, 553556.
McGrath,K.P., Fournier,M.J., Mason,T.L. and Tirrell,D.A. (1992) J. Am. Chem.
Soc., 114, 727733.
Nagarajan,V., Kamitori,S. and Okuyama,K. (1999) J. Biochem. (Tokyo), 125,
310318.
Noelken,M.E., Wisdom,B.J.J. and Hudson,B.G. (1981) Anal. Biochem., 110,
131136.
Obrecht,G., Lefe`vre,J.F. and Meyrueis,P. (1991) Fr. Pat. Appl. 91 16215.
Ohara,O. and Teraoka,H. (1987) FEBS Lett., 211, 7882.
Pollack,S.V. (1990) J. Dermatol. Surg. Oncol., 16, 957961.
Privalov,P.L. (1982) Adv. Protein Chem., 35, 1104.
Rao,K.P. (1995) J. Biomater. Sci. Polym. Ed., 7, 623645.
Reynolds,J.A. and Tanford,C. (1970) Proc. Natl Acad. Sci. USA, 66, 1002
1007.
Rossi,A., Zuccarello,L.V., Zanaboni,G., Monzani,E., Dyne,K.M., Cetta,G. and
Tenni,R. (1996) Biochemistry, 35, 60486057.
Saddler,J.M. and Horsey,P.J. (1987) Anaesthesia, 42, 9981004.
Smith,P.K., Krohn,R.I., Hermanson,G.T., Mallia,A.K., Gartner,F.H.,
Provenzano,M.D., Fujimoto,E.K., Goeke,N.M., Olson,B.J. and Klenk,D.C.
(1985) Anal. Biochem., 150, 7685.
Sreekrishna,K. and Kropp,K.E. (1996) In Wolf, K. (ed.), Nonconventional
Yeasts in Biotechnology. A Handbook. Springer, Berlin, pp. 203253.
Suzuki,Y., Ikeda,N., Kataoka,E. and Ohsuye,K. (2000) Biotechnol. Appl.
Biochem., 32, 5360.
Van Heerde,G.V., van Rijn,A.C., Bouwstra,J.B., de Wolf,F.A., Mooibroek,H.,
Werten,M.W.T., Wind,R.D. and van den Bosch,T.J. (1999) Eur. Pat. Appl.
0926543.
Vedvick,T., Buckholz,R.G., Engel,M., Urcan,M., Kinney,J., Provow,S.,
Siegel,R.S. and Thill,G.P. (1991) J. Ind. Microbiol., 7, 197201.
Wardi,A.H. and Michos,G.A. (1972) Anal. Biochem., 49, 607609.
Werten,M.W.T., van den Bosch,T.J., Wind,R.D., Mooibroek,H. and de
Wolf,F.A. (1999) Yeast, 15, 10871096.
Wilcox,P.E. (1967) Methods Enzymol., 11, 607617.
Zacharius,R.M., Zell,T.E., Morrison,J.H. and Woodlock,J.J. (1969) Anal.
Biochem., 30, 148152.
Received December 11, 2000; accepted April 23, 2001

454

Anda mungkin juga menyukai